Outliers Impact on Clustering. Sol & Nieve is a sporting good and outdoor gear retailer that operates in North America and Central America. In an attempt to characterize its stores (and re-assess Sol & Nieves supply chain operations), Gustavo Esposito is analyzing sales in 22 regions for its two primary product lines: sol (beach-oriented apparel) and nieve (mountain-oriented apparel). Gustavo has generated the following output for k-means clustering for k 5 2, 3, 4 (output reported in standardized units). Which value of k is the most appropriate for these data? How should Gustavo interpret the results to characterize the clusters? k 5 2 Inter-Cluster Distances Cluster 1 Cluster 2 Cluster 1 0 2.215 Cluster 2 2.215 0 Within-Cluster Summary Centroid (Original Units) Size Average Distance Sol Nieve Cluster 1 11 0.655 25.148 6.695 Cluster 2 11 0.685 6.340 25.276 k 5 3 Inter-Cluster Distances Cluster 1 Cluster 2 Cluster 3 Cluster 1 0 4.603 1.977 Cluster 2 4.603 0 3.076 Cluster 3 1.977 3.076 0 Within-Cluster Summary Centroid (Original Units) Size Average Distance Sol Nieve Cluster 1 11 0.655 25.148 6.695 Cluster 2 1 0 6.500 60.000 Cluster 3 10 0.154 6.324 22.932 k 5 4 Inter-Cluster Distances Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 1 0 4.458 3.060 6.064 Cluster 2 4.458 0 1.728 3.076 Cluster 3 3.060 1.728 0 4.457 Cluster 4 6.064 3.076 4.457 0 Within-Cluster Summary Centroid (Original Units) Size Average Distance Sol Nieve Cluster 1 1 0 60.000 6.500 Cluster 2 10 0.154 6.324 22.932 Cluster 3 10 0.120 20.099 6.715 Cluster 4 1 0 6.500 60.000